When companies talk about cloud computing, the benefits that immediately come up are cost efficiency, agility and scalability. But cloud infrastructure management does more than that. It not only speeds up application development but also encourages innovation and transforms the business culture, says Mai-Lan Tomsen Bukovec, vice-president and general manager of Amazon S3, a web service offered by Amazon Web Services (AWS).
She tells Enterprise that cloud computing products and services are popular because they address the core concepts of what application developers and businesses want — cost savings and agility. “With the pay-as-you-go model, you don’t have to provision all the hardware because you just use what you need. And if you don’t need it, you just release the instance [each time a computer programme runs].”
Nick Walton, managing director for Asean at AWS, says cloud offerings address the issue of overcapacity as businesses are able to match their capacity with the spikes in customer demand. When the capacity is no longer necessary, they may spin it back down and stop paying for it.
Walton points out that in the old world of on-premise infrastructure, the required capacity size has to be estimated and if a business errs on the side of caution, it ends up buying more than it needs and could be bogged down by the infrastructure. “But let’s say you have a very successful marketing campaign and the demand and traffic to your website increase. If you are not able to meet that capacity requirement, you end up with dissatisfied customers, which is counterproductive to your marketing activity,” he adds.
Bukovec says once businesses get started, however, they realise that the real benefit of cloud computing is that it shortens the time to market. For instance, Thailand’s stock exchange, which serves about 100,000 customers during peak periods, rebuilt its whole online trading system on the AWS platform and saved up to 30% in infrastructure costs.
“But the real value was being able to deliver the whole online infrastructure in just eight months. This is a story we hear over and over again and it does not matter which customer or region,” says Bukovec.
“They start because they want to save on costs. But they find that it also changes their development time, how they develop and the speed at which they can come up with new ideas, and run little experiments to see if they like it or not. This goes into the culture of the business and starts to be a very important foundation for digital transformation.”
Walton says AWS works proactively with its customers to help them spend as little as possible. “We have a tool called Trusted Advisor, which is an automated platform that helps customers recognise where they are wasting resources. So, if they are using more computing power or storage than they need, the tool will make recommendations. We send out millions of notifications every year and have seen customers save more than US$350 million annually.”
Access to reliable security tools
Bukovec says small and medium enterprises (SMEs) usually start adopting cloud computing by migrating their websites to the AWS platform. Over time, they move other workloads such as data processing.
“Raw storage and backup are other common scenarios and the SMEs [that migrate to the AWS platform] will immediately have access to the strongest security tools. This may not have been possible or economical if they managed it themselves,” she says.
Walton says AWS customers’ content and security requirements have been central to how the web service provider thinks about delivering its offerings. He points out that historically, encryption has been a complicated process, but AWS has a rich set of tools that make it much simpler.
“So, customer content and data are owned and controlled exclusively by them and they provide control over who has access to the content and data. We do not have access to the data. That is a key part of how we build trust with large enterprises that store and run critical applications on AWS today,” says Walton.
AWS has trained more than 25,000 people in Southeast Asia over the last seven years to help them build the kind of capabilities and skills required to benefit from cloud computing, according to Walton. “Attitudes are changing rapidly, and one of the things we have seen is that customers understand the level of security we have built into the platform and see it as an opportunity to be more secure than if they were using their own data centres and servers.”
Bukovec says when customers in the financial sector believe that AWS’ security is better than their own, it is very telling indeed. “Capital One’s chief investment officer said AWS runs its data centres more securely than it does. That is one of the reasons they moved to AWS. The regulatory body of the US stock exchanges is building all of its systems on AWS for the same reason. Deutsche Bank is another customer that has said AWS is safe enough for banks to use.”
The machine learning component
In her keynote speech at the Amazon Web Services Summit in Singapore last month, Bukovec said “builders” are all over organisations. It is no longer just the developers but also business and operations people as well as senior leaders.
“No longer can you expect that the only input to your applications is through a keyboard or mouse. Now, you have different types of input, ranging from voice to data sensors, with data streaming in from Internet of Things devices. Connectivity is everywhere,” she told the 6,000 participants at the summit.
This means enterprises that leverage cloud computing have to iterate very quickly, she pointed out, adding that it is no longer okay to do a deployment (of functions in applications) once a week or once a month. Deployments are now happening 10 to 50 times a day.
“You have to deploy at this speed to keep changing your customer application because today, user context is everything. The user experience is personalised, customised and deeply embedded in the users’ world, and your application has to understand that,” said Bukovec.
She notes during the interview that one of the trends in cloud infrastructure or architecture revolves around data and points out that the two aspects of data usage — insights and analytics — can be done through machine learning (ML) capabilities.
“We have a lot of ML going on at Amazon. And if you have billions of orders coming in, an application or a human being will find it very hard to go through every single one of them to identify the characteristics and patterns of the customers. So, ML is an incredibly useful technology and Amazon has been using it for 25 years,” says Bukovec.
“Amazon Prime Air (which is still in development) will use drones to deliver packages within 30 minutes of the order on the website. Meanwhile, Netflix says more than 75% of what is watched today is driven by recommendations. These are all driven by ML.”
She notes that targeted advertising is commonly done on AWS as it requires an incredible amount of computing capability. The visual heavy Pinterest, for example, uses ML tools on AWS to help classify the images.
“We have a customer in Southeast Asia called Shoppr, which analyses about 100 million images on social media every day and makes inferences based on which fashion brands are showing up in these images. It shares the data with the fashion brands themselves and makes recommendations,” says Bukovec.
Regardless of the type of industry, if there is a vast amount of data, businesses can use ML to understand the patterns and use this to power different types of applications, especially smarter applications, she adds. “What will happen is you will see how ML transforms the user experience into something more personalised and relevant to what customers really want.”